Bootstrap coav Yocto layer

[mbelluzzo]

This layer contains recipes to build and install coav-control and it's init script on a Yocto image.

This patch was tested on the Intel Aero Yocto build and due the python recipes provided by the version of meta-openembedded on the build, a custom configure step is needed in order to use python to build the Mavlink headers. Once meta-openembedded gets updated to a newer version, that task may be removed.

Signed-off-by: Murilo Belluzzo murilo.belluzzo@intel.com

58

[rchiossi]

LGTM, but I think this should be in it's own repo (that's the 'yocto way' of doing things). cc: @anselmolsm

[anselmolsm]

@rchiossi 's point makes sense. Submit this to github.com/intel-aero/ ?

[rchiossi]

This is for anyone working with yocto, not only for aero. It would be better to just give it its own repo (see soletta and aero itself for instance).

[rchiossi]

Once we have it on its own repo, we can send a pr to meta-aero to add it as default to aero's recipe.

[mbelluzzo]

I think its better to make it "public" once we hit milestone 3. To avoid frustrate expectations about the library. I don't mind to have it here first and move for a separated repo later because right now the usage is very specific.

Once we have the recipe for the library itself, I think we can change to a new repo and also submit the recipes to meta-intel-aero.

[anselmolsm]

This is for anyone working with yocto, not only for aero.

Just being pragmatic: Aero is the first platform where it's going to be tested and used. I don't see a random guy taking a random recipe to build that targeting a random drone platform, at least in the short term. Placing these recipes together with other Aero recipes is a benefit, IMO.

However, I'd first keep the focus on making the lib ready for prime time. Then we can revisit this discussion about moving the recipes.